Authors: Ana Petra Jukić, Ana Šelek, Marija Seder, Ivana Podnar Žarko

Published on: May 10, 2024

Impact Score: 7.4

Arxiv code: Arxiv:2405.06473

Summary

• What is new: A deep learning dual-model solution for autonomous driving combining braking and steering using modified NVIDIA’s PilotNet and MobileNet SSD.
• Why this is important: The need for efficient autonomous driving systems that can operate on devices with constrained resources.
• What the research proposes: Modifying the NVIDIA PilotNet model to reduce its resources requirement, making it more suitable for smaller devices while maintaining performance.
• Results: The modified PilotNet model shows similar autonomous driving performance to the original, with reduced model size and inference latency.

Technical Details

Technological frameworks used: Deep learning, NVIDIA’s PilotNet, MobileNet SSD

Models used: Modified PilotNet, MobileNet SSD

Data used: Single front-facing camera images

Potential Impact

Automotive industry, particularly manufacturers and tech companies focusing on autonomous driving technologies

Want to implement this idea in a business?

We have generated a startup concept here: StreamDrive AI.